Skip to content
Audio

vits-ru-mms-graphemes: A Russian text-to-speech model in ONNX format

1 min readupdated

This model is a variant of Meta's VITS architecture, based on the facebook/mms-tts-rus project and repackaged to work with the phoonnx ONNX library. It takes Russian text as input and generates high-quality audio (a WAV file). The model is convenient for integration into Python applications and voice assistants, including the OVOS (Open Voice Operating System) ecosystem. It will be useful for developers of voice bots and autonomous systems.

Key features

  • Based on the proven VITS architecture for speech synthesis
  • Full support for the Russian language and character-by-character text input
  • Optimized for running through the ONNX format using phoonnx
  • Supports integration into the OVOS voice engine

What you can do

  • Add local Russian speech synthesis to your Python script or bot
  • Configure the voice engine in the OVOS configuration file for a smart home

Example

import wave
from phoonnx.model_manager import TTSModelManager

manager = TTSModelManager()
manager.load()
manager.merge_default_voices()

voice = manager.voices["facebook/mms-tts-rus-Russian"].load()

with wave.open("output.wav", "wb") as wav_file:
    text = "Привет, это синтез речи."
    voice.synthesize_wav(text, wav_file)
Read moreDetails
Architecture: VITS
Base model: facebook/mms-tts-rus from Meta
Library: phoonnx (ONNX text-to-speech)
Text normalization: via the scriptconv converter script
License: cc-by-nc-4.0
Discuss Telegram 0 15

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.6 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # Phoonnx/vits-ru-mms-graphemes
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download Phoonnx/vits-ru-mms-graphemes --local-dir ~/models/vits-ru-mms-graphemes
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to vits-ru-mms-graphemes
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of vits-ru-mms-graphemes (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

Phoonnx/vits-ru-mms-graphemes
Author
Phoonnx
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

150

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments